Rain lashed against the window of Mark's cramped apartment, mirroring the chaotic churn in his stomach. He hadn't slept properly in days. The email, discovered accidentally, sat open on his laptop: a rejection from a local bakery for a "part-time baking assistant" position.

Across the room, David, his supposed “software architect” friend, was polishing his collection of antique model trains. The rhythmic swishing of the cloth was maddening.

"So," Mark began, his voice sounding thin, "the big project at Global Dynamics… how's that going?" He held his breath, waiting for the answer.

David glanced up, his face carefully neutral. “Oh, you know. Slow and steady. Lots of meetings.” He returned to his trains. Mark noticed his hands shaking.

Emotion: tense

Cluster: Fear / Anxiety
PC1 (Valence): -1.32 Negative
PC2 (Disposition): -0.53

Role in Research

This story is one of 1,000 stories generated for the emotion tense. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 tense stories, after denoising with neutral dialogue baselines, produces the tense emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the tense vector is projected through the unembedding matrix.

Promoted:
S0.213
😰0.191
느껴0.183
😣0.183
😖0.181
Suppressed:
a-0.260
de-0.258
la-0.256
happy-0.218
B-0.198